Collection of normalized and installable hunspell dictionaries.
- What is this?
- When should I use this?
- Install
- Use
- List of dictionaries
- Examples
- Types
- Contribute
- License
This monorepo is a bunch of scripts that crawls dictionaries from several sources, normalizes them, and packs them so that they can each be installed and used in one single way. Dictionaries are not maintained here but they are usable from here.
You can particularly use the packages here as a programmer when integrating with
other tools (such as nodehun, nspell) or when making
such tools.
In Node.js (version 12.20+, 14.14+, or 16.0+), install with npm:
npm install dictionary-en👉 Note: replace
enwith the language code you want.
⚠️ Important: this project itself is MIT, but eachindex.dicandindex.afffile still has its original license!
import dictionaryEn from 'dictionary-en'
dictionaryEn(function (error, en) {
if (error) throw error
console.log(en)
// To do: use `en` somehow
})Yields:
{dic: <Buffer>, aff: <Buffer>}👉 Note: preferred BCP-47 codes are used (according to Unicode CLDR). To illustrate, as American English and Brazilian Portuguese are the most common types of English and Portuguese respectively, they get the codes
enandpt.
In total 92 dictionaries are provided.
This example uses dictionary-en in combination with nspell.
Show install command for this example
npm install dictionary-en nspellimport dictionaryEn from 'dictionary-en'
import nspell from 'nspell'
dictionaryEn(function (error, en) {
if (error) throw error
const spell = nspell(en)
console.log(spell.correct('color'))
console.log(spell.correct('colour'))
})Yields:
true
falseThis example loads the index.dic and index.aff files located in
dictionary-hyw (Western Armenian) from a Node.js JavaScript module (ESM).
It uses a ponyfill (import-meta-resolve) for an
experimental Node API.
Show install command for this example
npm install dictionary-hyw import-meta-resolveimport fs from 'node:fs/promises'
import {resolve} from 'import-meta-resolve'
main()
async function main() {
const base = await resolve('dictionary-hyw', import.meta.url)
const dic = await fs.readFile(new URL('index.dic', base))
const aff = await fs.readFile(new URL('index.aff', base))
console.log(dic, aff)
}This example loads the index.dic and index.aff files located in
dictionary-tlh (Klingon) from a Node.js CommonJS script (CJS).
Show install command for this example
npm install dictionary-tlhconst fs = require('node:fs')
const path = require('node:path')
main()
async function main() {
const base = require.resolve('dictionary-tlh')
const dic = await fs.readFile(path.join(base, 'index.dic'))
const aff = await fs.readFile(path.join(base, 'index.aff'))
console.log(dic, aff)
}Follow these steps to use a dictionary on macOS:
- Navigate to the dictionary you want on GitHub, such as
dictionaries/$code(replace$codewith the language code you want) - Download the
index.affandindex.dicfiles (i.e., open them, right-click “Raw”, and “download linked files”) - Rename the download files to
$code.affand$code.dic - Move
$code.affand$code.dicinto the folder~/Library/Spelling/ - Go to System Preferences > Keyboard > Text > Spelling and
select your added language (it should come with the
(Library)suffix and is situated at the bottom)
The dictionaries are typed with TypeScript.
Yes please! See How to Contribute to Open Source.
To build this project, on macOS, you at least need to install:
- wget:
brew install wget(crawling) - hunspell:
brew install hunspell(many dictionaries) - sed:
brew install gnu-sed(crawling, many dictionaries) - coreutils:
brew install coreutils(many dictionaries) - ispell:
brew install ispell(German)
👉 Note: sed and the GNU replacements should be setup in PATH to overwrite macOS defaults.
Dictionaries are not maintained here. Report problems upstream.
Dictionaries are not maintained here. Most languages have a small community or institute that maintains a dictionary, and they often do so on GitHub or similar. Please ask in the issues to request that such a dictionary is included here.
👉 Note: acceptable dictionaries must:
- have a significant affix file (not just a
.dicfile)- have an open source license
- have recent contributions
See license files in each dictionary for the licensing of index.dic and
index.aff files.